Decentralized Orchestration of Open Services- Achieving High Scalability and Reliability with Continuation-Passing Messaging

Abstract

An ever-increasing number of web applications are providing open services to a wide range of applications. Whilst traditional centralized approaches to services orchestration are successful for enterprise service-oriented systems, they are subject to serious limitations for orchestrating the wider range of open services. Dealing with these limitations calls for decentralized approaches. However, decentralized approaches are themselves faced with a number of challenges, including the possibility of loss of dynamic run-time states that are spread over the distributed environment. This thesis presents a fully decentralized approach to orchestration of open services. Our flow-aware dynamic replication scheme supports both exceptional handling, failure of orchestration agents and recovers from fail situations. During execution, open services are conducted by a network of orchestration agents which collectively orchestrate open services using continuation-passing messaging. Our performance study showed that decentralized orchestration improves the scalability and enhances the reliability of open services. Our orchestration approach has a clear performance advantage over traditional centralized orchestration as well as over the current practice of web mashups where application servers themselves conduct the execution of the composition of open web services. Finally, in our empirical study we presented the overhead of the replication approach for services orchestration.